Appraising, processing, and providing access to email in contemporary literary archives

Abstract

The email of contemporary literary figures is ripe for research by scholars, and of broad interest to the general public, but can also present many challenges to cultural memory institutions that seek to appraise, process and provide access to this rich archival material. This article explores how five institutions across the United States, United Kingdom and New Zealand are using ePADD, free and open source software developed by Stanford Libraries that incorporates artificial intelligence and machine learning to help meet these challenges for email in contemporary English-language literary collections. Authors and institutions represented include British poet Wendy Cope (The British Library), British novelist Ian McEwan (Harry Ransom Center, University of Texas at Austin), British Indian novelist and essayist Salman Rushdie (Emory University), American poet Robert Creeley (Stanford University) and New Zealand poet and critic Ian Wedde (National Library of New Zealand). The use cases are followed by a discussion identifying lessons learned and areas for further research.

Finding light in dark archives: using AI to connect context and content in email

“…Some map easily to pre-digital archival practices. For instance, an email corpus belonging to a single individual can be treated as a collection of correspondence (Schneider et al 2019). Respondents can be identified, rights requested and granted or denied, and access managed according to updated, but otherwise known practices.…”

Natural language processing and machine learning as practical toolsets for archival processing

“…For this overview, we have focused on the appraisal module. See Schneider et al (2019) for an overview of the software, along with current user documentation [20]. Among available NLP software dedicated to archival processes, it is the most mature, so it is worth exploring its functionality in some detail.…”

Receptor: A Platform for Exploring Latent Relations in Sensitive Documents

“…However, differently from the system of Abualsaud et al [1], Receptor focuses on indicating sensitivities by automatically finding latent relations between entities and events that appear in sensitive or non-sensitive documents. Recently, some efforts have been initiated to build tools to support finding sensitive information in a collection most notably, ePADD [4] and BulkReviewer. 2 ePADD [4] is an open-source software that supports the discovery Figure 1: Layered system architecture and appraisal of email archives.…”

“…Recently, some efforts have been initiated to build tools to support finding sensitive information in a collection most notably, ePADD [4] and BulkReviewer. 2 ePADD [4] is an open-source software that supports the discovery Figure 1: Layered system architecture and appraisal of email archives. ePADD incorporates natural language processing (NLP) techniques, including named entity recognition (NER), to assist users in searching within email collections.…”

“…Receptor is implemented in python. spaCy 4 is used for all of the NLP operations and the web-interface is implemented using the Django 5 framework. The graph database is supported by Neo4j.…”